From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mailgw1.uni-kl.de (mailgw1.uni-kl.de [IPv6:2001:638:208:120::220]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 9D0073B29D for ; Wed, 3 Mar 2021 02:44:13 -0500 (EST) Received: from sushi.unix-ag.uni-kl.de (sushi.unix-ag.uni-kl.de [IPv6:2001:638:208:ef34:0:ff:fe00:65]) by mailgw1.uni-kl.de (8.14.4/8.14.4/Debian-8+deb8u2) with ESMTP id 1237iBE2062025 (version=TLSv1/SSLv3 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT) for ; Wed, 3 Mar 2021 08:44:11 +0100 Received: from sushi.unix-ag.uni-kl.de (ip6-localhost [IPv6:::1]) by sushi.unix-ag.uni-kl.de (8.14.4/8.14.4/Debian-4+deb7u1) with ESMTP id 1237iAxs010841 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO) for ; Wed, 3 Mar 2021 08:44:10 +0100 Received: (from auerswal@localhost) by sushi.unix-ag.uni-kl.de (8.14.4/8.14.4/Submit) id 1237iAgJ010839 for bloat@lists.bufferbloat.net; Wed, 3 Mar 2021 08:44:10 +0100 Date: Wed, 3 Mar 2021 08:44:10 +0100 From: Erik Auerswald To: bloat@lists.bufferbloat.net Message-ID: <20210303074410.GA9809@unix-ag.uni-kl.de> References: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) X-Spam-Status: No, hits=-0.999, tests=ALL_TRUSTED=-1,URIBL_BLOCKED=0.001 X-Spam-Score: (-0.999) X-Spam-Flag: NO Subject: Re: [Bloat] receiver side ledbat? X-BeenThere: bloat@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: General list for discussing Bufferbloat List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 03 Mar 2021 07:44:13 -0000 Hi, On Tue, Mar 02, 2021 at 05:21:25PM -0800, Dave Taht wrote: > I always kind of thought that clamping the receiver side window more > dynamically would help a lot on cellular networks. I concur and therefore disabled TCP window scaling when using Internet access via cellular networks. It is not sufficient for EDGE and worse, but OK for 3G and fine for LTE in my experience. > https://tools.ietf.org/html/draft-irtf-iccrg-rledbat-01 I really like the idea of having a way to tell the sender side to slow down, since they would otherwise fill the huge buffers in cellular networks (I have measured round-trip-times of up to 4 minutes over cellular networks). The classical thinking that it is fine to send whatever the receiving host can handle no longer holds in our bufferbloated networks. Thanks, Erik -- And I always like seeing NSA-designed cryptography [...]. It's like examining alien technology. -- Bruce Schneier